Anda di halaman 1dari 16

Design of Unlined and Lined Pressure Tunnels

R. P. Benson
Abstract--The paper covers the engineering requirements for designing a safe and economicalpressure tunnel. Design requirements and criteria set out include the geological aspects of positioning and alignment; protection against failure by hydraulic jacking; determination of unlined versus lining requirements; selection of temporary and final linings utilizing rockbolts, shotcrete, or concrete; design of the steel-lined portion for length, load-sharing, external water pressure, buckling, installation and grouting; design of debris traps and plugs; and requirements for watering up and dewatering.
R6sum~--Les tunnels ~ puissance doivent de plus en plus f onctionner sons des pressions de plus en plus grandes. Dans certain cos, sp~cialement pour des projets ~ haute pression, des d~faillances se sont produites alors que tous les criti,res traditionnels de conception avaient ~t~ respectS, et des rt~thodes ar~lior~es furent alors n~cessaires. L' auteur eonsidkre des nouveaux critkres de conception pour les revStements de tunnel, et en particulier pour les eas des hautes pressions. Les criti,res inclus sont: la position et l'alignement du tunnel, les protections eontre les pouss~es hydrauliques et les pouss6es verticales, la s~lection des revStements flnaux, la conception de la section ~ rev~tement en acier, et les aspects op~rationnels.

Introduction
ower tunnels are required to convey water from the power intake to a powerhouse, which may be on the surface or underground. Such tunnels may begin in soil or rock, and often pass through a variety of materials and geologic conditions before reaching the powerhouse. Their prime responsibility is to convey the water safely throughout the life of the project, without detrimental effects on the surroundings. Such effects may include excessive leakage from the tunnel, instability of surface soil or rock resulting from seepage, saturation and softening of agricultural land, and pollution of groundwater and surface streams due to organic content of the tunnel water. These effects can be controlled by careful positioning of the tunnel, and by selecting the appropriate lining and treatment for the various parts of the tunnel. It is essential to understand the geologic conditions along the tunnel alignment, relative to the hydraulic forces that will be applied during operation of the tunnel. Adequate investigation techniques and tests now exist that will define the geologic and geotechnical conditions. There are also appropriate materials with which to line the necessary parts of the tunnel, and to treat the material surrounding the tunnel. However, it is necessary that the conditions be investigated and that the designs be established by geologists and engineers experienced Present address: Dr R. P. Benson, President, Klohn Leonoff Ltd, 10180 Shellbridge Way, Richmond, British Columbia, VBX 2W7, Canada. This paper is reprinted with permission from Canadian Tunnelling 1987/88 with some modifications.

in engineering geology and applied rock mechanics. Power tunnels are being called upon to perform under increasingly higher heads. Numerous existing hydroelectric plants have heads beyond 1000 m, and projects are now reaching upwards of 1500 m. Although high-head tunnels have been built in the past, such tunnels essentially have been steel-lined throughout the high-head portion. Modern practice for such projects is to provide shorter steel liners, subjecting long unlined portions of the tunnels to high hydrostatic pressures. The behavior of rock and the tunnel linings under these high hydraulic pressures is difficult to assess, and the necessary design techniques and safety factors must be judged carefully. In some cases, especially for highhead projects, failure has occurred, even though traditional criteria for design have been met. It is apparent, therefore, that these traditional criteria are not fully adequate and that improved design methods are necessary. What has constituted safe practice in the past for relatively low-head projects--practice based on simplified or empirical design criteria---can no longer be considered acceptable.

Positioning and Alignment of the Tunnel


Although selection of the tunnel alignment is governed primarily by economics, a number of other factors must be taken into account in establishing the optimum alignment. Such factors may be divided into two categories: primary and secondary. The primary factors that are likely to have a dominant role in overall economics include: Schedule for project completion.
0886-7798/89 $3.00 + 0.00 Maxwell Pergamon Macmillan plc

Available contractors and equipment. Difficult geological conditions. Many tunnels are designed without appropriate consideration and ranking of these primary factors, which strongly affect the final cost. When they arise during the course of construction, the result is costly overruns or project delays. Many power tunnels are short, less than 1 km; however, a large percentage can be up to 10 km long. Many are significantly longer, up to 30 or 40 km, and may take a very long time to complete. However, the schedule can be improved, for example, by providing additional access adits; by specifying a certain manner of construction (mole vs drill-and-blast); or by ensuring that a majority of the tunnel is driven in sound rock via a longer route where progress will be rapid and reduced support and lining will be required. Therefore, the project schedule must be carefully considered and an understanding obtained of potential costs or savings if the schedule is delayed, or achieved ahead of time. With such information in hand, the designer can then consider the correct options. The availability of experienced contractors and modern equipment can be a major concern, especially in underdeveloped countries. Owners may be constrained to require local or national contractors, or may not be willing to allow or risk the use of high-speed tunnelling machines. In the latter case, the use of standard drill-and-blast technology can result in a longer construction schedule. Appropriate access adits, a longer schedule, or special high-speed drill-and-blast working methods must then be specified. The possibility of encountering difficult geological conditions must be

Tunnelling and UndergroundSpace Technology, Vol. 4, No. 2, pp. 155-170, 1 9 8 9 .


Printed in Great Britain.

] 55

carefully assessed, and avoided if possible, by selecting an alternate route, even if it appears more costly at first glance. These conditions can often be determined through preliminary evaluations by engineering geologists or engineers experienced in difficult tunnelling. In some cases, such an evaluation may require drilling or other field investigations before even a preliminary geologic assessment can be properly undertaken. Nevertheless, such an assessment is mandatory if alternative alignments are to be ranked appropriately and rationally. There are many secondary factors that affect the selection of the tunnel alignment. Some of the most common and most important of these include the following: Grade. Access requirements. Prevention of hydraulic jacking or uplift. Position of water table. Temporary support. Final lining. Ventilation. Drainage. The tunnel must be maintained below the hydraulic grade line for all modes of power plant operation, including hydraulic transients, to prevent negative pressures in the tunnel. This requires assessment of head losses by friction along the tunnel. The degree of friction, of course, varies with the type of lining or, in the case of an unlined tunnel, with the roughness of rock walls. Head losses may be determined by the Darcy-Weisbach formula or the empirical Manning equation. The appropriate roughness coefficient must be selected. The Manning formula, developed for the open channel flow and subsequently adapted for pipe flow, is generally applicable for conduits with diameters greater than 2 m, whereas the Darcy-Weisbach formula is theoretically suitable for a wider range of roughness values. Selecting the appropriate coefficients requires careful judgment.. Table 1 provides a summary of roughness coefficients

for various types of linings. These coefficients cover a wide range and are taken from examples provided in literature (Soivik 1984) and from experience with a variety of tunnels. In selecting the precise coefficient to be used, a wide range of variables for the actual tunnel must be considered. Positioning of the tunnel to ensure submergence beneath the hydraulic grade line should be made with conservatively selected low-friction coefficients to ensure against the possibility of entrapping air in the tunnel arch. However, for determination of head losses for calculation of energy losses, more realistic values must be used. Given the primary requirement of linking the intake to the powerhouse, innumerable alignments may be chosen. Technical and economic analyses must choose among factors, e.g. highlevel vs low-level alignments, the use of shafts vs inclined sections, all predicated on the primary requirement of maintaining containment of hydraulic pressures without failure. Figure 1 shows a number of commonly used variations. Each of these arrangements has advantages and disadvantages that

must be compared bet0re the most economical arrangement is selected. A number of key areas that often are not given adequate consideration are: Construction Access. The number of access adits to the power tunnel required to meet the overall schedule is often misjudged, especially if driving rates are not as assumed. Adding extra adits during construction may disrupt the basic design, and may compound contractual problems and costs. In addition, extra adits must be suitably plugged for operation. Intermediate access adits also may be provided to handle suspected delays resulting from poor geology, where no amount of exploration can reasonably be done in advance. Access for Steel Liner Installation. This requirement may be overlooked, especially for underground powerhouse arrangements. Installing steel liners is best done as a separate operation, with as much independent access as possible, especially if the steel liner extends upward into the shaft. Prevention of Hydraulic Jacking. The entire tunnel, including the shaft and the surge tank, must be set deeply

L
A. SHORT-COUPt.ED-UNDERGROUND POWERHOUSE B. HIGH LEVEL P~)WER TUNNEL WITH INCLINED SHAFT

C. LOW LEVEL POWER TUNNEL ANO INTAKE SHAFT

D. HIGIt LEVEL AND LOW LEVEL POWER TUNNEL WITH SHAFT

E. 81.OPING POWER TUNNEL

F. ~ LEVEL POWER TUNNEL WITH SURFACE PENSTOCK

NOTES : Q stmOE FACILITIESNOT SHOWN SINGLETUNNELONLy - DISTRI~JI[IONSYSTEMSNOT 81tOWN SURFACEOit UNOEI~QInO~J(~IO POWEfllIOUSE A8 811OWN

Figure 1. Various power tunnel schemes.

Table 1. Comparable roughness coefficients for Manning and Darcy--Weisbach.


Comparable Darcy-Weklbach"f'forvarying diameters Uning Unlined (drill and blast) Unlined (tunnel boring machine) Shotcrete (drill and blast) Concrete Steel Manning's"W 0.025-0.040 0.016-0.022 0.018-0.025 0.012-0.016 0.010-0.014 2.5 m
0.057-0.147

5 m
0.046-0.117

7.5 rn
0.040-0.102

10 m
0.036-0.093

0.023-0.044 0.030-0.057 0.013-0.023 0,0092-0.018

0.019-0.035 0.024-0.046 0.010-0.019 0.0073-0.014

0.016-0.031 0.021-0.040 0.0092-0.016 0.0064-0.012

0.015-0.028 0.049-0.036 0.0083-0.015 0.0058-0.011

Note: Higher values should be used when overbreak areas are used to calculate velocities.

156

TUNNELLING AND UNDERGROUNDSPACETECHNOLOGY

Volume 4, Number 2, 1989

enough within the rock mass to ensure that adequate in-situ compressive stress is available to prevent hydraulic jacking. Overlooking this design problem can lead to disastrous and very cosily consequences. A detailed discussion of how to avoid this problem is given below. In addition to the items above, the selection of temporary support and final lining, and the handling of ventilation and drainage during construction are important elements in establishing the position and alignment of the tunnel. These aspects are discussed more fully below.

Protection Against Hydraulic Jacking and Uplift


General Concepts
The concept of hydraulic jacking, or uplift, is well known. This effect can develop if water pressures imposed within a rock mass are greater than the in-situ compressive stress. Depending upon the deformability of the rock mass and the area over which the hydraulic pressures act, existing joints can be opened. This action may result in jacking of a large mass of rock away from the tunnel, resulting in excessive leakage and large-scale landslides or instability. This effect has occurred many times in the history of civil engineering, and has had disastrous consequences for both surface and underground hydroelectric facilities. Hydraulic jacking will occur in any direction where movement of rock masses can develop due to a lack of adequate compressive in-situ stress. Thus, vertical lifting of horizontally bedded rock,jacking or hoisting of rock masses towards valley walls, jacking of rock blocks into adjacent underground openings, or opening of fractures in a compressible rock mass can occur. Many authors have reported on cases of rock lifting and failure of the tunnel, with resultant seepages, landslides and even flooding of underground powerhouses. Brekke and Ripley (1987) recently completed an intensive review of the subject, and have listed a number of failures as long ago as 1920. When such failures occur, it often takes many months to diagnose the problem and complete repairs. The cost of repairs and loss of revenue from power sales may amount to tens of millions of dollars, causing considerable distress to inhabitants due to the shortage of power. Therefore, a careful and conservative design to prevent problems of this nature is warranted. The cost of additional necessary steel liner or reinforced concrete liner to ensure a conservative design is well justified when compared with project delays and lost revenue, the cost of actual repairs, and the mental anguish of having to redo an unsafe design.

For many years, unlined tunnels were located by ensuring that the weight of the rock and, in some cases, the soil vertically above the tunnel was at least equal to the static water head. This relationship was simplified to provide rock cover of at least half the head, because the density of rock is normally at least twice that of water. This simplified approach was generally adequate for horizontal surfaces, although for rock having specific gravity lower than 2.0, the factor of safety would be lower than unity. Many failures have occurred as a result of a deficiency in vertical cover. However, failures also have occurred where the tunnels approached valleys, and either frontal or lateral cover was low, or where geologic conditions resulted in low stresses. Particularly, the problem of low lateral stresses near valleys has resulted in many failures due to an inadequate length of steel liner. Some designers who recognized this problem either set the tunnel back from the valley or used reinforced concrete to span the section. However, reinforced concrete is not necessarily an appropriate solution in preventing hydraulic jacking. In some cases, it only affects the timing of the problem, as the pressures tend to build more slowly because of the decreased seepage through the reinforced liner. Various designers and investigators recognized this problem, and took steps to either measure the rock stresses or estimate the probable stress levels by stress analysis. Thus, photoelasticity and, ultimately, finite element methods were used to assess stress condition around valleys. In addition, methods of stress measurement by overcoring and by hydraulic jacking have been developed that allow designers to position the tunnel rationally, or to design a suitable lining to resist hydraulic forces if the tunnel cannot be safely located. Similarly, other design methods, e.g. grouting, and pressure relief and drainage, have been used to limit seepage pressures and thereby prevent hydraulic jacking. However, such methods are not considered to be a safe approach and cannot be recommended. Hydraulic jacking can be properly prevented only by a steel liner or a reinforced concrete liner that accepts the full internal pressure without contribution from the rock. Designing to control hydraulic fracturing by grouting and/or drainage should only be used where potential failure can be tolerated, or where a problem has arisen that cannot reasonably be solved by a more direct approach. However, such measures must always be taken in conjunction with sound geological interpretation, to ensure that the design measures are compatible with the behavior of the various geological materials when sub-

jected to high pressure seepage water. Material boundaries, probable stresses, permeabilities and deformability must be determined by appropriate geologic and testing methods.

Measurement of In-Situ Stress


In-situ rock stresses may be measured by a variety of methods using overcoring techniques to obtain data on the three-dimensional stress field. Hydraulic fracturing tests conducted in drill holes are used principally to obtain a measure of the minimum principal stress in the rock mass; however, they also can be used to measure the three-dimensional field (Cornet 1986). These tests must be performed and interpreted carefully to ensure that in-situ stress of the rock mass is being measured, rather than some characteristic of the pumping system. Thus, the equipment must have a pumping capacity that will ensure response of a reasonable mass of rock, such that its inherent permeability does not affect the results. It appears also that the shut-in pressure is too unreliable a criterion to adopt, and assessment of in-situ stress should be based on the reopening pressures measured by repeatable tests. If the tests are done from an underground opening, they must be at a distance 1.5 to 2.0 times the diameter of the opening to ensure that the in-situ stress has not been affected by the presence of the opening. Both methods of measuring in-situ stress are important, depending on the type of problem faced by the designer. However, the hydraulic jacking test should always be performed, as it simulates the actual effects that will be imposed on the rock by the tunnel water. It is also a direct measure of the minimum principal stress, and requires neither mathematical calculation nor application of theoretical formulae. It is important that the drill holes used for hydraulic jacking tests cross alljoint systems, especially through-going or master joints, and those that are subparallel to valley walls. Where the minimum principal stress is not the vertical stress, and where deformable rock or shear zones exist, positioning a tunnel to meet only vertical stress criteria may not be adequate. Figure 2 shows a power tunnel positioned beneath a horizontal surface cut transversely by a major fault with low lateral stress, and a low deformation modulus. Lateral hydraulic jacking can open vertical fractures, allowing excessive seepage to the surface. For this condition, siting of the tunnel based only on cover criteria could be unsafe. Knowledge of the complete stress field, the rock modulus, and the position of the water table is necessary. For deformed rock masses that have

Volume 4, Number 2, 1989

TUNNELLINGANDUNDERGROUNDSPACE TECHNOLOGY 157

LEAKA@E AT ~ A C E LEVEL-H

Where there is no soil cover, o r where the soil could be removed by landslides, the equation reduces to:

1.3Hw
gr = T h e above equations provide a factor of safety of 1.3 against uplift on horizontal planes, such as bedding planes or joint combinations. In the event that the lateral stresses are lower than the vertical stresses, consideration must be given to hydraulic jacking against vertical planes (joints or faults), particularly where deformable rock exists (Fig. 2). It is necessary to be cautious in designing projects to be built on valley slopes, where seepage from the tunnel at the end of the steel liner could induce landslides, even when special drainage is provided. In this instance, it is prudent to assume that the cover must be provided by the rock portion only. Pressure tunnels positioned near slopes or valley walls demand detailed consideration of the stress environment. Particularly in ridges, noses, or near valley walls, stress relief of the rock mass occurs. This phenomenon can be seen in the existence of open jointing near valley or cliffs; this jointing may extend tens to hundreds of meters back from high cliffs. The loss of stress is magnified in the upper portions of the valley wall or cliff, in contrast to the base of the valley, where stress intensification commonly exists. Broch (1984) recommends that the topography be diagrammatically corrected to match the overall topographic contours of the surrounding landscape. This procedure must be done both longitudinally and laterally to account for irregular topography beside the tunnel (Fig. 3). This method will ensure safe positioning of the sloping portion of the pressure tunnel and of the surge shaft, if one is required. Nevertheless, even when these apparently conservative rules are adopted, hydraulic jacking tests should be conducted at the critical points along the tunnel. The critical points are the end of the steel liner, the upper elbow, the upper portion of the surge shaft, and major slope changes of the valley wall, as shown in Fig. 3. An appropriate equation follows for obtaining a suitable position for a pressure tunnel where the rock conditions are favorable, i.e. generally hard with high stiffness (modules o f deformation), and where the topography is uniform with near-horizontal or gentle slopes. This equation will provide a factor of safety of 1.3 against uplift or hydraulic jacking. For areas of complex geology or topography, where locally low stresses may occur, and especially for high head plants, the probable stress patterns should be evaluated

.......

MODULUSR O C K

titlt.tit;,iz~=, '

!z!:]t!;!l!l!l:at.

POWERHOUSE J

\ ]

TUNNEL PROFILE CONTAINING LOW MODULUS ZONE

60,000~" ~
~u 5 0 . 0 0 0

ASSUMED ~'v IONS ( ~'N ~ CONDIT AT 0.$


H 500m

.oooo , , , , _ _ _
20,000 Io,ooo /On Zr~._
I
I ~. I 3

,w

i 4

ELASTIC COMPRESSION LOWMODULUS OF ZONE c m *


Figure 2. Hydraulic jacking in a vertical plane.

zones or beds of stiff and flexible material, the stress field may be highly variable. Thus, it is possible for the water pressure to dilate fractures in the low-stress deformable rock surrounding the tunnel, and to force an open pathway to a nearby powerhouse or access adit-----even if the overall rock cover is adequate. This effect has been noted in deformed granite masses, and in sandstone/siltstone sequences where stresses measured by hydraulic jacking tests were very low in the soft deformable sandstones, b u t were higher in the more brittle sihstone. Care must be taken to obtain representative tests of those rocks where hydraulic jacking in compressible rock can occur during operation, such as short seepage paths to nearby u n d e r g r o u n d openings, or a highly pervious zone exiting from the rock mass. Similarly, it argues for a sound understanding o f the geology and the material properties of the materials to be penetrated.

construction, it is p r u d e n t to measure the in-situ stress as soon as is practicable. In addition to checking the safety of the final design, measuring the in-situ stress will allow optimization of various design elements---such as the precise location of the tunnel, surge tank, sloping penstock, elbows, length o f steel liner--and, thus, optimization of the design. For a pressure tunnel positioned beneath a near-horizontal surface, traditional criteria generally are adequate. It is necessary to calculate the overlying weight o f soil and rock separately, due to differences in the density of the two materials. This factor can be significant in areas o f deep tropical weathering. T h e following equations govern these calculations:
=

1.3 H w - H , y~
Y,

H,

Design Guidelines
In the event that data on in-situ rock stress are not available, simple design rules can be followed that will ensure safety. However, if the project proceeds to final design and

= height o f the rock above the tunnel (m); % = density o f rock (t/ma); H s = height of soil above the tunnel (m); % = density of soil (rimS); and H~ = maximum static head (m).

158

TUNNELLING AND UNDERGROUNDSPACE TECHNOLOGY

Volume 4, Number 2, 1989

/ ~ . . . . . - - -

ACTUAL CONTOURS

/.///

ACTUAL SLOPE /CORRECTED SUOPE

F~ESSU~[ TUmlEL/"

"

_
I-

STEEL LINER

Figure 3. Broch correction for topography.

in more detail by a method such as finite elements to check the validity o f the equation.

H,.-

1.3Hw
Cos 0

o m m e n d e d design line, shown on the figure, can be udlized for valley slopes up to 60 . For slopes steeper than 60 , special analysis o f the probable in-situ stress would be required, especially to

determine the minor principal stresses. T h e design line ensures a factor of safety o f 1.3 against static conditions. T h e p o r d o n o f the factor o f safety above unity will normally provide protection against surge and unexpected geologic factors. As noted above, the in-situ stresses should be measured at key points by hydraulic jacking tests, at the earliest opportunity. Many options are available for bringing the water to the powerhouse. For control o f individual units, individual tunnels may be necessary. Two power tunnels may serve a n u m b e r o f units by utilizing bifurcations, or a single power tunnel may feed a distributor, with separate tunnels leading to the powerhouse. Each of these layouts imposes different hydraulic pressures on the various tunnels and their linings for the various operational conditions. Guidelines for siting the pressure tunnel are the same for u n d e r g r o u n d powerhouses and surface powerhouses. T h e sloping portion o f an unlined or concrete-lined tunnel must be positioned properly, relative to the valley wall. H o w e v e r , for an u n d e r g r o u n d powerhouse it is also necessary to consider the effect o f the tunnel on the stability o f the u n d e r g r o u n d openings. T h e length of the steel liner required to maintain high-pressure water a safe distance from the other openings becomes an important consideration. Individual pressure tunnels must be separated by an adequate distance, normally 1.5 to.2 times their diameter, to ensure pillar stability. In the event that one or more tunnels remain operational while others are unwatered, full static head must be considered to act on the other tunnels. For the powerhouse or other galleries, stability of the rock surrounding the openings must be maintained, as seepage water can cause hydraulic jacking if the in-situ rock stress is too low (Fig. 5). In this

H r = required rock cover (m);


Hw = static head (m); % = density o f rock (t/mS); and 0 = slope angle. Note that the surge shaft must receive special consideration because the u p p e r portion may extend above the line established by the Broch correction (Fig. 3). Thus, unless it is steel lined, the surge shaft should be set far enough back into the slope to ensure that the in-situ rock stress at the top o f the shaft is adequate. Brekke and Ripley (1987) have compiled a n u m b e r o f case histories where excessive seepage has resulted from hydraulic fracturing or lifting. These cases are compared with a number o f similar cases where embedment o f the tunnel is deeper, and where no trouble occurred, as shown on Fig. 4. A rec-

I,I

1.0

0,9 0.B II "r ~. 0.7 w 0.6 O

u
a O O& 6 n U 4

DE$1G~ILINE FOil "l'r 2.75 gm/cc


N O & A A

I E A O (m) OIO0

NO OR 8MAU. LEAKAOE O

81GI~ICANI" LEAKA~ I

I 0 0 - 200
200 - 300

0
A

A Q

0.5

0.Z-

0.!

0o

I
-

O O O O 0 Ve

~W V O a

lwA m
eg

Imlrl , idlpold Itglm OrokkI iml aoB~r l I N T ) ~tqnd IIiif IIIIOI! I I i I O ,

I
I0"

I
20"

I
30"

I
40"

I
IO*

I
~O*

SLOPE ANGLE - deg.

Figure 4. Recommended design line to prevent hydraulic jacking.

Volume 4, Number 2, 1989

TUNNELLING AND UNDERGROUND SPACE TECHNOLOGY

159

MAJOR PRINCIPAL STRE$S~

MINOR ORtNCIPAL

,E.sTo:KJ

_1

END ST~L ~ , K \
LINER ~
\

CROSS SECTION

NORMAL ~
TO A e - ~ ' " /

~. "m~ CLOSURE ERROR

~m.._L~. ,~ .m~
PENb'rOCK

. /FAILURE PLANES

11N , , ~ { "

NORMALFOR~
SHEAR ~

# //

..-SHEAR ~qCE
FORCE >.~

t~/ l / -/$HE~R =',

~'i~'--" m~.3SUR
L r...,._ ~,

a ~.
,~""~O

~- O " F o .a 2L
u.! ~

/ PRESSURE ~-"--""'L.O.J RELIEF

,~rrwou'r

N ton V E C T O R ANALYSIS OF POTENTIAL BLOCK

20" 40' 60" 8C~' ANGLEOF PLANE WWH HORIZONTAL S U M M A R Y OF STABILITY ANALYSIS

Figure 5. Rock jacking at underground opening.


event, the steel liner may have to be lengthened, and special drainage galleries and drill holes installed to relieve high-pressure water. In addition to the consideration of in-situ rock stresses and the imposed hydraulic loads, a safe design requires that the hydraulic gradients towards openings be controlled, especially if erodible z o n e s intersect both openings. Appropriate filtering of erodible zones intersected by drainage holes may be necessary. T h e distance between the end o f the steel liner and the opening must maintain hydraulic gradients at acceptable levels. Guidelines are given below, u n d e r "Steel-Lined Section." prevent hydraulic jacking. For this reason, water h a m m e r transients are not required to meet a normal factor

of safety; however, a factor o[ safety for normal operational pressure surges is required. Recommended factors of safety for various hydraulic conditions are given in Table 2. A safe and conservative design that prevents hydraulic jacking or lifting requires that the steel liner be carried an adequate distance to ensure that in-situ stress criteria are met. In specific cases, a reinforced concrete liner can be used where the liner carries the full internal pressure without contribution from the rock. In this case the concrete contributes no tensile capacity. When it is considered necessary to design for hydraulic jacking, it is essential to limit the area around the tunnel where hydraulic jacking can be allowed, either with pressure relief and drainage induced by drill holes or nearby openings. This technique has been used to limit the growth of hydraulic jacking in rock masses and to close cracks in concrete gravity dams. Alternatively, grouting a zone around the tunnel to limit the amount of seepage into the rock mass and prevent the build-up of pore pressures beyond the grouted zone is conceptually valid. However, it is essential to ensure that the rock outside the grouted zone is permeable or drained, and that large areas are not subjected to the jacking forces of high-pressure water. It can also be argued that the hydraulic pressure causing hydraulic jacking is the pressure on the outer edge of the concrete liner, decreased due to the head loss that develops as the water seeps through cracks in the liner. This is not considered to be a safe and valid design approach, because it is impossible to predict accurately the width and distribution of cracks in a concrete liner. Cracks may be highly variable, depending upon concrete placement and curing, variable mixes, and the variable deformability of rock under pressure. Thus, the hydraulic pressure that must be resisted by the

Table 2. Recommended factors of safety against hydraulic jacking or uplift.


Normal operating Design condition Lifting of rock above horizontal unlined or concrete lined tunnel. Along sloping portion near valleys, and at end of steel liner, with proper allowance for slope, topography, and possible landslides removing soil cover. Static Surge Water hammer

1.3"

1.1

N/A

1.3

1.1

N/A

Recommended Factors of Safety


Prevention o f hydraulic jacking may be considered as simply ensuring that the hydraulic pressure within the tunnel is always less than the rock stress, or that the time of application of the hydraulic stress is too short to

* May be reduced to 1.2 if geological conditions are well-known.

160

TUNNELLING AND UNDERGROUNDSPACETECHNOLOGY

Volume 4, N u m b e r 2, 1989

rock to prevent hydraulic jacking must be equal to the internal pressure in the tunnel, including surge conditions. Limiting hydraulic jacking by using pressure relief or grouting should not be used as a general design approach. However, if such measures are necessary because of special conditions, they should be performed with great caution. For success, it is necessary to understand the geologic conditions and the variations in permeability of the rock to ensure that high-pressure seepage paths do not develop that bypass the pressure relief or grouting systems installed. In general, this technique should be viewed as a last-resort measure, and not as recommended practice.

Selection of the Final Lining


Selection of the final lining for a pressure tunnel is a process that begins in the design stage but does not end until construction is complete and the geological conditions are known in detail. During the tendering stage, the liner design must be considered to be preliminary. Then, as construction p.roceeds and the geological condiuons become known, the lining may be modified to suit the actual conditions. This procedure requires well structured and flexible specifications and contract documents that allow design modifications without unfair penalty to the contractor or the owner. There is wide divergence on the use of linings by various designers throughout the world. Some favor completely lined tunnels on the basis of ensuring long-term performance without the need for any maintenance work during operation. Others have championed the advantages of unlined or partly lined tunnels, with acceptance o[ local fallouts provided that they do not prejudice operations. Between these limits are many options that attempt to optimize cost and performance. There are no common guidelines for selection of the final liner. Three basic factors influence the selection of the lining: (1) Achieving acceptable head loss in the conduit. (2) Preventing excessive leakage either by seepage or hydraulic fracturin_ g. (3) Ensuring long-term stability during watering up, operation and unwatering. Each of these factors is discussed in detail below.

a smaller but smoother lined tunnel may be the least costly. For tunnels in excess of approximately 6 km, the schedule often can be controlled by the excavation and lining requirements of the power tunnel. Thus, selection of a smaller lined tunnel vs a larger unlined tunnel may be an important decision. Head loss through a conduit is principally a function of the wall roughness, the tunnel diameter and the water velocity. As a result, hydraulic equivalence can be obtained between largerdiameter unlined tunnels vs smaller lined tunnels of greater hydraulic efficiency. When small tunnels (2-3 m in diameter) are used, there is a greater need for a smooth lining to maintain acceptable head losses. However, as the tunnel diameter increases, the wall roughness has less effect on head loss, and equivalence is achieved through small diametral changes. These factors must be assessed in terms of excavation and support costs for both the temporary and final linings, and the schedule advantages of unlined tunnels where the host rock permits an unlined tunnel. The advantages of a tunnel boring machine, which can achieve a smoothness almost equivalent to concrete, also must be considered. Selecting the most suitable lining for the aspect of head loss is a complex but standard matter of project economics.

Leakage Control
Excessive leakage from pressure tunnels can occur in two ways: first, by hydraulic jacking; and, second, if the rock is pervious and the internal pressure exceeds the external groundwater pressure. Methods of dealing with hydraulic jacking have been discussed above. The problem of pervious rock is more difficult, as

there are numerous methods that can be considered, and the possibility of misjudgment increases accordingly. Some seepage loss may be allowed, depending upon the quantity and value of available water and the probable effect of seepage on the stability of the terrain and its effect on the environment. Therefore, the long-term value of the water vs the cost of limiting the seepage must be compared. I f possible, of course, the tunnel should be placed in such a way that the hydraulic grade line is below the water table. Because determination of the permanent groundwater table is very important, experienced geologists and groundwater hydrologists should be involved in this task. Mapping of all springs and points of periodic seepage, as well as drill-hole piezometric data, are essential in such an evaluation. If this information is not obtained, a conservative design approach is necessary. Excessive leakage from the unlined tunnel can occur if the natural groundwater surface is lower than the head developed by the internal pressure, and if the rock is pervious. O f course, a low water table may exist even in relatively impervious rocks, which are found in many very dry areas of the world. The various cases are shown in Fig. 6. Leakage for unlined tunnels can be estimated by standard flow nets if investigations have provided appropriate data along the length of the tunnel. Such data normally are obtained by geologic mapping, drilling and in-hole permeability testing. Subjective judgment is almost the only guide for determining how much investigative work is necessary, and this judgment must rest with an experienced engineer and/or hydrogeologist. There are three basic types of lin-

-~"
CASE A NO LEAKAGE

II

T"

........

~ -J

GASE B -

LIMITED OR NO LEAKAGE

Head Loss
Acceptable head loss along the pressure tunnel is a matter of overall project economics, which, in some cases, may involve using an unlined tunnel and achieving an earlier on-line schedule. Alternatively, where the schedule is controlled by other project elements,
GA8E

G -

PO881BLE HIGH LEAKAGE

CASE

D -

PO88iBLE

HIGH LEAKAGE

Nw

STATIC

HEAO

Figure 6. Assessmentof leakage conditions (unlined, shotcrete, or unreinforced concrete).

Volume 4, Number 2, 1989

TUNNELLINGAND UNDERGROUNDSPACETECHNOLOGY 161

ings: impervious, pervious and sentipervious. In practical terms, the most c o m m o n impervious lining is steel, placed with a concrete or concrete/ grout surround. A composite liner can be used, in which the steel is thin and an inside layer o f concrete provides buckling resistance against external pressure. In recent years various types of impervious lining have been used, including bitumen-coated copper, sprayed rubber, and plastic, normally with a concrete liner placed inside to prevent buckling. However, very little performance data exist for such linings, and their design and construction are complicated, as they must handle both internal and external pressure. Concrete and shotcrete applied without reinforcement must be considered to be pervious because they are characterized by locally pervious zones due to placement imperfections, or shrinkage cracks that have occurred during curing. They also are easily cracked under internal pressure in deformable rock zones. Unfortunately, there is a c o m m o n misconception that concrete and shotcrete linings are impervious. Both concrete and shotcrete can be reinforced to act as semi-pervious linings, utilizing enough reinforcing to distribute and control cracks to a specific width. If a semi-pervious liner is required, it is necessary to provide reinforcing both radially and longitudinally, and to control concrete mixes and placement carefully. Leakage rates through semi-pervious, cracked concrete linings can be reasonably estimated using equations for flow between plates. Figure 7 shows the relation for laminar flow and turbulent flow, which allows calculation o f leakage from a cracked concrete or shotcrete liner (Sawatsky 1986). Note that seepage is basically a function of the crack width to the third power, the head, and the permeability of the surrounding material. T h e relations also show that if the permeability o f the surrounding material is relatively low, the leakage is almost the same for one crack as for many. T h e graph indicates that if the surrounding rock is pervious, significant seepage can occur unless the crack widths are very small. Generally, a rock permeability less than 10-~ cm/sec would allow very minor losses. Effective, comprehensively applied grouting of pervious rock can reduce permeability to between 10--4 and 10--5 cm/sec. I f the host material is relatively permeable, leakage control depends on reinforcement and/or grouting. Although standard structural methods are available for design of the reinforcing steel, appropriate determination o f the deformation modulus is required for the rock in order to calculate the deformation o f the

:[
o Jr__

NOTES

: lOOm

='/
/

1
-

I I

r~NEL
~ h PROFILE

HOMOGENEOUS ROCK (IJnllmm P l r m e o b i l l l y , K )

1, NET HEAD

lOOm

~ ~

SINGLE -f~SECTION *hl~n*s'

2. FLOW PATH LENGTH 1HROUGR ROCK lOOm 3. TUNNEL LINER I ~ C K N E S S 4OOmm

IOO

E,o

js
I OOOOI I
,,,,

R CRACK ONLY I ossumet no bockplessufe lyom rock }


e

IIII 0.OI DISCHARGE 0,1 - malsec

I per

I I Illll lO0m

1 Itlllll I0

I IIliil I00

O.0OI

length

Figure 7. Leakagefrom cracked concrete liner.


lining. Various methods are available for this determination (see discussion under "Stability," below). T h e number of cracks that will occur in an elastic, homogeneous, circular liner principally depends on the diameter and liner thickness, and generally is in the range o f 10-20. Reinforcement ensures that the cracks will be distributed, and the crack width can then be calculated with confidence. Cracking of the concrete lining also develops for other reasons, including placement conditions, cold joints, temperature effects, excessively rich or wet mixes, and shrinkage. Longitudinal shrinkage often results in regularly spaced transverse cracks separated by 6-10 m, depending on the diameter of the tunnel. Such cracking can be hairline or up to 3-4 mm, depending on the concrete mix and curing properties. Grouting to reduce permeability around a concrete liner can be successful, provided that carefully controlled techniques and pressures are used. Generally, cement grouts with various additives to control penetration, thixotropy, setting time, and strength are available. Silica fume, used in shotcrete, also has been used in grout to increase strength and decrease the water/cement ratio. Grouting should be done in rings, normally with six to eight holes per ring, after the concrete or shotcrete is placed to allow higher pressures to be used. Pressures must be controlled to prevent damage to the lining, but must be high enough to ensure penetration. Consolidation grouting pressures generally should reach at least 75% of the internal pressure, although a maxim u m 15 bars normally would be adequate. Stage grouting, as shown on Fig. 8, has been successful where grout loss has to be controlled and where high pressures are necessary. New chemical grouts that can penetrate joints and rocks with a semi-pervious matrix also have been utilized.

Stability
T h e final lining selected must ensure adequate stability of the tunnel throughout the life of the project. For many designers, this means that no rockfalls or falls of shotcrete or concrete should occur. Such an approach requires a continuous concrete liner, or very highquality rock with extensive rock bolting and shotcrete for an unlined tunnel. However, other designers accept considerably less support and lining with some minor falls in the tunnel, provided they do not hamper operation or cause a significant energy loss. The differences in cost and construction schedule between these two design approaches can be very great, and significant economic benefits are possible by selecting the approach that allows for minor rockfalls, in combination with periodic inspection and maintenance. To ensure stability, the lining designer must consider: Erosion of rock or joint filling by pressurized water. Rock support, temporary and final. Hydraulic pressures during watering, operation, and unwatering.

162

TUNNELLING AND UNDERGROUNDSPACETECHNOLOGY

Volume 4, Number 2, 1989

CONSOLIDATION GROUT HOLES't~'~-~.~.~" ~

T --~

I
/
/

I [ GROUT HOU=.'~ "~.:.~, :.: ~'>d/CO.CRETE [~'...x.,.,.,.,.,.,.,.~ -'~'~i .-'~ LINER I

tCO

\
!

covered with shotcrete or concrete. I f shotcrete is used, it must adhere properly, as hydraulic gradients can be very high during unwatering. Controlled unwatering, where pressures are reduced over a period of days, is desirable to reduce seepage pressures and external forces on the lining.
T e m p o r a r y and final rock support

SEQUENCE

1. STAGE 1 - GONTACT GROUTING LOW PRESSURE 2. STAGE 2 - LOW PRESSURE GROUTING TO PROVIDE CONTAINMENT OF STAGE S GROUT 3. STAGE 3 - HIGH PRESSURE TO REDUCE PERMEABILITY LIMIT TO 75% OF Pi

Figure 8. Liner grouting.

Erosion

The velocity of water in unlined pressure tunnels normally does not exceed 4 rrdsec, and often is in the range of 2-3 m/sec. Such velocities can cause a progressive erosion of weak rock, resulting in rockfalls that can reduce the capacity of the tunnel or, in the extreme, entirely block the flow. The erosion may occur in soft rocks, shear or fault zones, or in blocky rock containing thin, clayey or silty seams and veins. Although such erosion may be slow initially, it can progress to involve large blocks of rock and falls measuring thousands of cubic meters. However, such zones can be treated during construction to ensure trouble-free operation. The project locations requiring protection must be identified during excavation by experienced engineers or geologists during the geologic mapping of the tunnel, and inspected during walk-through surveys by experienced personnel. After these features have been identified, the necessary treatment can be specified. Generally such treatment involves hand excavating to a depth equal to the width of the seam, and covering the area with fiber or mesh reinforced shotcrete. For particularly poor zones, stitching

with criss-crossed, short, grouted rock bolts is adequate unless the feature requires covering with a concrete liner. Such decisions can only be made by e x p e r i e n c e d personnel. It also is wise to make one final, thorough walk-through inspection with design and construction personnel just as the tunnel is being completed, to ensure that no zones have been missed. Installation of reinforced shotcrete or concrete can be done either at the time of driving or when excavation is complete. Excavation of seams to allow treatment may be necessary. Although shotcrete normally can be used, it must be reinforced and anchored to the rock because the fine-grained materials can prevent adhesion of the shotcrete. Where adhesion is poor, fiber-reinforced shotcrete is not recommended; mesh-reinforced shotcrete is superior. I f the zone is broad and very deformable, it may be necessary to cover it with a concrete liner. Weak, erodible rock or fine-grained material from seams or faults can be piped into the tunnel through cracks, joints or drainage holes, especially during unwatering, when external pressures around the tunnel are high. Pipeable rock must be completely

Rock reinforcement is required on a temporary basis during tunnel driving, and for permanent support during operation of the power tunnel. The degree and type of support for the two conditions can be very similar, especially for hard, massive rock where no final lining is required, and where only rock bolts and shotcrete are required for both temporary and final support. In contrast, the temporary support required for tunnel driving in weak, strongly fractured, or highly erodible rock, may be modest, whereas extensive support or lining may be necessary for operation. Thus, it may be necessary to follow excavation and temporary support with subsequent extensive additional support, i.e. up to a complete concrete lining. The temporary and final support requirements invariably are established by experienced tunnel engineers during driving of the tunnel. However, the general support requirements must be determined during design so that appropriate specifications and construction methodology can be adopted to ensure an economical project and the minimum construction schedule. Various classification systems exist for assessing temporary support. In 1946, Terzaghi developed a rock classification system for estimating the loads to be supported by steel sets in the tunnels. Since then, a number of classification systems have been developed on the basis of numerous case histories. Some of the more common classification systems include: Terzaghi's Rock Load Classification for Steel Arch Supported Tunnels, 1946. Deere's RQD and Merritt's Method, 1972. Wickham's Rock Structure Rating, 1972. Barton's Norwegian Geotechnical Institute (NGI) Index Classification, 1976. Bienawski's CSIR Method, 1976. Although these systems are helpful in both design and construction planning, they do not address all variables; nor are they uniformly applicable to all types of rock and various conditions of rock quality. They must be applied by people having good judgment based on experience with various geologic conditions. Walia (1985) has presented a good summary of the appficability of these systems, as summarized in Table 3.

Volume 4, Number 2, 1989

TUNNELLINGAND UNDERGROUND SPACETECHNOLOGY 163

Table 3. Evaluation of rock mass classification systen~c


Geologic Conditions Discontinuities closely spaced or weak planes X Sheared or crushed rock otherwise rehealed X Orientation of one or more significant discontinuity

Main factors affecting stability Rock relaxation (Time dependent)

Sound maealve rock X

Discontinuities moderately spaced X

In-situ stress - Rockpopping Conditions - Squeezing conditions(time dependent) - Slaking (due to loss of moisture)
- Swelling

X X

X X

X X

X O

O O

O
O

O
O

O
O

O
O

O
O

Groundwater - Inflow volume - Erosion susceptibility


- Solutioning

X O
O

X O
O

X O
O

X O
O

X O
O

- Hydrostatic pressure Gas - Pressure


- Potential

O
O

O
O

O
O

O
O

O
O

explosion X - Adequately addressed. O - Inadequately addressed or not addressed.


Selecting the type and amount of support for the final lining is a complex problem. However, based on the relatively simple classification systems, a combination of support may be employed, subject to individual preferences. Nevertheless, the support requirements in all the classification systems are related fundamentally to the strength o f the rock, the size o f the tunnel, and the specific geologic conditions. Figure 9 and Table 4 summarize these factors and identify the probable type o f support that is appropriate for various types o f rock. T h e information shown is intended for use in pressure tunnels with diameters ranging from 2 to 7 m, which is the normal range for most projects. It must be emphasized that the information is given as a guideline only, and that any specific tunnel could have conditions for which the measures shown may be inappropriate. Nonetheless, the data are based on a n u m b e r o f cases and, used as a guideline, have been helpful for both design and construction. solutioning, the tunnel can remain largely or fully unlined. Special zones or areas o f weakened rock can be treated with grouted rock bolts and shotcrete. An estimate of the rockbolts and shotcrete required can utilize the classification systems discussed above. T h e underlying design philosophy is that minor rockfalls can occur, but that these falls will not lead to larger major falls that could reduce tunnel capacity. T h e minor falls will either lie on the invert, or be moved by flow to a rocktrap set near the end o f the tunnel, upstream, o f the steel lining. Rockbolts are used either sporadically, to pin individual loose blocks, or on a pattern basis to provide an interactive support for unlined tunnels. Many types o f bolts are available and various theories for rock bolt designs have been developed. Although both active and passive anchorage rockbolts systems have been used for tunnel support, in recent years the tendency for modestsized tunnels has been to use a passive system o f grouted but untensioned anchors. In hard rock, where plucking or solutioning will not occur between bolts, this method is economical and effective. Following are some additional specific design points for unlined, shotcretelined, and concrete-lined tunnels.
Shotcrete fining

Unlined tunnels Provided that the rock is hard and durable and not susceptible to

Shotcrete is an effective way of improving the stability of tunnels. Used in conjunction with rock bolts or passive grouted anchors, as in the New Austrian Tunnelling Method (NATM), a highly adaptable support system can be developed to meet conditions observed in the field during tunnel driving. With the newer additions, such as fiber-reinforcement and silica-fume, the strength and deformability can be varied to meet special requirements. Shotcrete can be used to improve the smoothness o f tunnels; a Mannings "n" value between 0.018 and 0.025 can be achieved, depending on the thickness and contouring achieved: Generally, shotcrete 2-5 cm thick has no effective strength; however, when it is applied in thicknesses o f I0 cm or greater, design methods similar to those for concrete can be utilized. Shotcrete can be used effectively in most rock conditions. However, for claystones or weak, poorly cemented rock, shotcrete may not adhere properly or may fail easily at the contact. For such conditions,

164

TUNNELLINGAND UNDERGROUNDSPACETECHNOLOGY

Volume 4, Number 2, 1989

:o
60

~,

~Q~II~ :
UIG Powerhouse or Large Tunnel c I n ~ r c ~ T U t ~ ; r n s l ' t e Tunnel

l
I

i~ 40 '
50

~ i

--

30

///.
/
/

/
o

.,
~,o~

J
.

subsequent layers. As many as four layers have been used in situations where concrete placement could not be done effectively. Single or double bars can be shotcreted, and triangular, tied bundles also have been effectively placed. As reinforcing is increased, reducing the maximum particle size of the shotcrete is advantageous.

/
I

Concrete linings
Concrete linings represent the most effective method of ensuring stability while achieving a hydraulically efficient tunnel. Because placement is normally done behind smooth steel forms, Mannings "n" values on the order of 0.012-0.016 are achieved. Construction methods have been developed that can provide homogeneous concrete, with minor imperfections such as cold joints, thermal cracking and honeycombing. Naturally, good mix designs and construction techniques are essential. Nevertheless, because of the inevitable imperfections, and because of the variable deformability of most rocks, linings cannot be considered impervious. All the evidence shows that the lining does not prevent transfer of pore pressures into the surrounding rock. Such evidence includes cracks observed to be filled with leaves and fragments upon unwatering; rises in the groundwater table, with development of surface leaks; and instrumented tunnels where piezometers have been installed in the host rock. In addition, cracking of both reinforced and non-reinforced liners, caused by shrinkage during curing, occurs to some degree with all linings. Concrete liners must be considered by the designer to be pervious or semipervious membranes. However, concrete can be reinforced to impede flow and, depending on the permeability of the surrounding rock, pore pressures may then be controlled to a specific zone around a tunnel. This concept has been discussed further above, under "Leakage Control." Because of the associated costs a n d increased construction schedule, selection of concrete as a support medium is a last resort. Nevertheless, in cases where other support measures will not be effective, e.g. in squeezing ground, where extensive overbreak has occurred, or in very deformable rock under internal hydraulic loading, concrete is an excellent solution. The lining must be designed for three conditions: (1) To support the external rock and water load that will be imposed during operation. In specific cases, swelling rocks may induce loading. (2) To ensure integrity of the liner under internal pressure, e.g. when seepage outflow is not of concern and concrete cracking without dislodging pieces is acceptable.

o/
/o
ti /

*~,~

" /

/
..* w'-

oO,,"
//
. J 7

./
/,4't ~O
.~??

I
_1

7
/

I 0

2 $
I0 IS 20 25

$0

SPAN

(m)

CLASSIFICATION NO SUPPORT MINOR SUPPORT

BASIC FINAL LINING UNLINED UNLINED

ADDITIONAL. SUPPORT MEASURES -UNTENSIONED GROUTED ANCHORS LOCALLY.

-UNTENSIONED GROUTED ROCK ANCHORS AS NECESSARY. - UNTENSIONED GROUTED "PATTERN ROCK ANCHORS LOCALLY - LOCAL SHOTCRETE. -ADDITIONAL REINFORCED SHOTCRETE. - TENSIONED GROUTED BOLTS OR TENDONS. - LOCAL CONCRETE. - CONSOLIDATION GROUTING AS NECESSARY. - STEEL RIBS AS NECESSARY. - REINFORCED AS NECESSARY. - PATTERN CONSOLIDATION GROUTING. - CIRCULAR RIBS AND PRIMARY CONCRETE LINING. - ADOITIONAL REINFORCING AS NECESSARY. - SPECIAL GROUTING.

MODERATE SUPPORT

$HOTCRETE AND PATTERN BOLTS

HEAVY SUPPORT

CONCRETE HORSESHOE OR INVERTED U CIRCULAR CONCRETE

vERY HEAVY SUPPORT

Figure 9. Rock support guideline.

interaction with the rock, achieved with mesh-reinforced shotcrete and pins, is not necessarily achieved with fiber-reinforced shotcrete. Therefore, the most effective shotcrete method must be selected. Shotcrete is a very adaptable material when the tunnel is subjected to internal pressures. Although shotcrete must be treated as a pervious material, crack distribution due to diametrical expansion occurs without concentration of cracks. It appears that adhesion to individual rock blocks is greater than the adhesion of concrete to rock. The rock blocks act independently, allowing cracking of the shotcrete to occur at block boundaries. Thus, when cracks concentrate in concrete liners, they are usually well-distributed in shotcrete liners unless major movement occurs. This factor modifies the approach to reinforcement design. Shotcrete also performs very well

during unwatering and allows dissipation of external water pressures without serious spalling, even though the strength as a membrane may be entirely inadequate theoretically. It appears that external pressures are relieved through the numerous but invisible cracks that have developed at joints and seams. Shotcrete has been observed to perform very well under all conditions where adhesion is good, even when repeated rapid unwatering has been performed. Thus, pressure relief is not necessary; and as long as good shotcreting practices are maintained, resulting in well-bonded shotcrete, good performance during unwatering can be expected. Reinforcement of shotcrete varies from fiber-reinforcement or welded wire fabric to layers of heavy reinforcing steel. With careful application, individual layers of reinforcing can be placed, shotcreted, and followed by

Volume 4, Number 2, 1989

TUNNELLINGAND UNDERGROUNDSPACE TECHNOLOGY 165

Table 4. Factors for selection of final lining.

Factor
Deformation modulus Geological favorability (faults, discontinuities, joint filling, orientation) Unconfined compressive strength Tunnel diameter Rock load on support* Rock mass permeability Rock mass quality (Q)* Rock strength/cover ratio * Based on NGI classification system

Unlined with rock bolts


>20 O00kg/cm 2 Favorable

Shotcrete lined
7000 - 20 000 Favorable to unfavorable 300 - 700

Concrete lined
<7000 Unfavorable

>700kg/cm 2

<300

Intensity of support increases with increasing diameter <0.4kg/crn2 <lxl0-Scm/s >20 >15 0.4 - 0.9 l x l 0 - S - l x l 0 -4 4 - 20 5 - 15 >0.9 >1x10-4 <4 <5

Note: The above factors are given as a general guideline only. To determine the final lining requirements, consideration of the cumulative effect of the
factors is necessary.

Table 5. Acceptable crack widths for concrete liners.

establish locations concrete-lined.

that

should

be

Design condition
External pressure Internal pressure Seepage control

Acceptable crack width


Not applicable 3 mm As low as 0.3 mm

Comments
Concentrated local cracks Distributed cracks See "Leakage Control"

SteeI-Uned Section

Length
T h e length of the steel liner must be carried to a point that satisfies two condititons: (1) hydraulic containment, and (2) acceptable hydraulic gradients. T h e requirement for hydraulic containment, or prevention o f hydraulic jacking, has been discussed above, under "Protection Against Hydraulic Jacking and Uplift." After this condition has been satisfied, it is necessary to ensure that the hydraulic gradient from the end o f the liner to the nearest exit point is sufficiently low to prevent instability o f soil or rock at the point of exit. Such instability can develop in the following ways, which have led to past failures on specific projects: Uplift beneath impervious soil or rock layers overlying more pervious rock. Erosion and piping of soil at the g r o u n d surface, or at pervious layers within the soil. Erosion and piping o f a fault or shear zone exiting at the ground surface, or at an u n d e r g r o u n d opening. Hydraulic jacking of rock blocks into u n d e r g r o u n d openings. An effective design against these conditions, as discussed above, is to prevent the development of excessive water pressures by installing suitable

(3) T o limit seepage outflow by limiting cracking to a tolerable amount (see "Leakage Control," above). T h e thickness o f the lining required depends on the size o f tunnel and the hydraulic forces that will be applied. Generally, thicknesses less than 0.2 m are difficult to place, and thicknesses greater than 0.7 m are rare. For the majority o f cases, where concrete is required for stability because o f external loading, reinforcement is not necessary. However, unusual cases such as squeezing rock or very high external water pressures may require reinforcing. Reinforcing is not required for the internal pressure condition unless severe cracking can occur, resulting in dislodging concrete pieces, especially during unwatering. Under internal pressure, deformations can

be allowed that would result in a few concentrated cracks of up to 3 mm before reinforcement would be necessary. I f reinforcement is considered necessary, it should be designed to limit cracking to 1 m m in width, with cracks distributed around the liner. Singh et al. (1988) have presented a simple method for reasonably estimating the cracking that can occur in concrete liners. Design methods are also available to calculate the deformations and load sharing a m o n g the rock, concrete and the necessary reinforcing steel (Hendron et al. 1987). Selecting the areas where a concrete lining should be installed requires experienced judgement. Following an assessment of geological factors, calculations o f possible cracking and/or reinforcement, the designers and construction personnel should inspect and

166

TUNNELLING AND UNDERGROUNDSPACETECHNOLOGY

Volume 4, Number 2, 1989

drainage. Provided that the high hydraulic pressures are dissipated in the mass rock near the end of the steel liner, no damage can occur. However, it is essential to understand the geologic conditions and the relative permeability of the various materials in designing the drainage system. The use of flow nets to study the flow regime and to assess the requirements for drainage is very helpful. There are no hard and fast design rules for acceptable hydraulic gradients in the rock mass; however, those shown in Table 6 are recommended for cases where the cover is adequate to prevent hydraulic fracturing or lifting. In conjunction with acceptable hydraulic gradients, drainage systems utilizing drill holes and separate drainage tunnels are the most effective method of maintaining stability or obtaining the shortest possible liner. Drill holes must be oriented to cut across the fracture systems, and must be spaced closely enough to ensure interaction. They should be drilled well below penstock or tunnel levels. Generally, holes spaced at about 3 m intervals are necessary, although more widely spaced holes can be effective in pervious rocks. Interaction between holes can be checked by pumping tests. Piezometers should be installed to measure the effectiveness of the drainage curtain during operation. If necessary, additional drainage can be implemented. If erodible material is encountered, drainage holes must be filtered to prevent piping. Control of water pressures and hydraulic gradients are important in the safe design of standard hydroelectric projects. For pumped storage projects, daily operations result in cyclic loadings and pressure changes. It is important to bear this in mind during the design of a pumped storage scheme, where more conservative design measures and increased factors of safety may be appropriate.

Table 6. Acceptable hydraulic gradients for steel liner length.

Type of material
Massive hard rock, widely jointed Hard to moderately hard, moderately jointed Moderate to weak rock, moderately jointed Weak, closely jointed or sheared Very weak, possibly erodible

Acceptable hydraulic gradients


10 - 15 8 - 10 5 - 7 3 - 5 Less than 3 plus appropriate filtering

maintain tensile stress, and variations in moduli among these materials. Depending upon the design criteria used to establish the requirements of the steel liner, these equations can be utilized to optimize liner thickness. The finite element technique is a more flexible and realistic method because it can be adapted to suit the actual shape of the tunnel and to include the existing stress field in a mass rock. As with the modified Lam6 equations, compatability of deformation across material boundaries is required. Figure 10 shows the results of a typical analysis for a cross-section with a variety of design variables that can be altered to study their effect. To obtain deformation moduli for the rock, suitable tests--such as plate jack, radial jack or pressurized chambers---can be utilized. However, these tests are costly and should be used only when significant economic advantage can be obtained by limiting liner thickness. For example, in many cases the

external pressure and designs to prevent buckling will override the internal pressure criteria in determining steel thickness. Other testing, such as hammer seismic or petite seismic methods, is useful to correlate between specific test stations and various rock types along the tunnel. Correlations among test methods, rock types, rock quality and deformation modulus are now more common in the literature. Because such correlations are adaptable to other sites, costly tests may not be required. The depth of the weakened zone around the tunnel is variable, depending on the quality of rock and the method of excavation. Machine-bored tunnels, of course, are much less damaged, and the low modulus layer generally is restricted to less than 0.3 m. For blasting, the low-modulus zone can be as deep as 2-3 m, but is generally less than 1 m. Seismic tests are useful in delineating the depth of this zone and its variation along the tunnel.

Internal P r e s s u r e
Given the rapidly expanding body of experience in treatment methods and evaluating rock properties for engineering purposes, the design of pressure conduits in rock in recent years has tended toward the rational optimization of steel and concrete liner requirements. Design considerations relate essentially to an assessment of the interaction of the steel, concrete and rock with respect to hydraulic forces that tend to cause instability or overstressing of these elements. Calculations normally have been based on modifications to the classic equations for thick-walled cylinders developed by Lain6 (1852). Generally, the Lain6 equations are modified to account for inclusion of a thin steel liner, the inability of concrete and/or rock to
/ "~/

DESIGN VARIABLE ~..4..---I " - " ~ t .~ ~ 11~8~ IpOCK " ~ .'4 :~ . ~..i-~ ~ " - - ~ "" ~ - ~ ' -t~t e
. - ~,~" ~ , ' a ~ ~ . . ~

-I- ~ #.,
~. -.

I. ROCK I~OOUt.U8 2. LAYERED ROCK 8YSTE M 3. STEEL AND O O N ~ T E LINER THIC NE88 s. cnAc:x,=o MNO t m c n A c x e a hOCK
,

/,',"

.~," - , ; " -" . 9 " - ~ ; : : . . ., ,.,

"~.'-'," ' ~J~'~"

"-,~.:. ",~-. \ ' , ,~, ~ , ~;, ''< " ~ , " "

,..: . I , "7"~'tt~ i, /

, .
~'Z:. :-..-.

.~:;" "~

7,

.OOT.

PIESllU~E8

/ t : ,t. ~ b 7
t l ~ I "l-~i:-"l I-;I ~

HOOP ITRtEBI

T~7"~,.,, i \ \
~.~i~-.-It.~-~.~/l ' ~# /''" /

i. I " i"

MODUI.U8 OF I I E E L - I I I I O i p l l -'OOOpiIITANGENiIALLY)

l."
~" .\

~ll,l,
"' ~

ll~."+;:'.~.

"'-~
"""?'.V
""-

i'_'.
"~ll/~"~" "

.1~ I,.. ~Z
" J l ' ~ I .i

l I 1400ULU, OF ROCK LAYEIq ,


MODULUS OF RO~" LAYE~!

',.i

"il \t" i

--

,~;';:'~,'~;'1i"
..oo,=,,.

!..

'

<"

TYPICAL STRES8 PLOT BY COMPUTER FOR RAOIALLY CRACKED CONCRETE

Figure 10. Finite element model of steel and conerete-lined tunnel.

Volume 4, Number 2, 1989

TUNNELLINGAND UNDERGROUNDSPACETECHNOLOGY 167

Low-modulus rock also exists near the toe o f slopes in river valleys, due to stress relaxation during the geologic processes forming the valley. In addition, the local mass rock near large u n d e r g r o u n d openings is weakened and stress-relieved by excavation of the opening. T h e liner must be designed to accept full internal pressure at such locations, with an appropriate transition to a point where the rock can sustain the load. Selection of the required length and transition depends on geologic evaluation of the rock. Where the rock is designed to accept load, design criteria for load sharing commonly require that the rock share half the load. This is most applicable to hard rocks, where creep is minor. However, in softer rocks, where creep is o f concern, the portion of rock sharing should be decreased accordingly. T h e allowable stress in the steel is normally limited to 50% o f the yield stress, and may be 35--40% o f the ultimate stress, depending on the type o f steel. It is general practice to ensure that the allowable stress does not exceed 80% of the yield, assuming no rock support, in the event that weak zones in the rock yield to equal the unrestrained expansion of the liner. T h e steel liner is normally expected to be impervious. However, for very high-head plants, leakage past the grout plugs has developed due to elongation o f the holes under the very high uniaxial stress, as unequal stress concentrations occur around the hole. Recent plug designs include backing plates and epoxy around the inner portion of the plug, ensuring a seal under internal pressure.

External Pressure
The steel liner is subjected to external pressures for the following conditions: Case A: Grouting. Case B: External groundwater pressures during construction, prior to watering. Case C: External groundwater pressures during unwatering following long-term operation. (In the case of rapid unwatering, many liners also are subject to partial vacuum, which acts in addition to the external pressure to instigate buckling.) Case D: External groundwater, plus possible excess pore pressure due to induced penstock pressures, or pressures from nearby operating penstocks. Over the last several decades, numerous investigators have developed theories predicting the onset o f buckling for the actual conditions under which steel liners are required to operate. T h e various theories attempt to relate measurable aspects of fabrication inaccuracies

to the types of buckling failure that can occur. Although no one theory fits all of the cases where failure has occurred, some of the parameters that might have contributed to failure have been quantified. The equations of Amstutz (1970) are the most suitable as a guide to specifying steel liner thicknesses. T h e factors having the most effect on the ability of a given liner to resist buckling are: The fabrication tolerances from which, in practice, the theoretically circular liner departs. The concrete mixes and methods of placing the backfill concrete between the steel liner and the rock. T h e ability o f the rock to recover from loading. The temperatures of the air and water to which the inside surface of the liner will be subject during critical conditions. An attempt should be made to estimate the effects of all of these factors before a realistic appraisal of the probable resistance to buckling of a given liner in a given situation is made. Fabrication tolerances should be specified realistically, bearing in mind the limitations imposed by available methods, materials and machinery. Particular attention should be paid to sudden changes in section that may occur at joints between plates; deviations from uniform curvature that may occur at plate ends, which are often bent rather than rolled; and deviations from circular, which usually occur as a result of inadequate support during installation and subsequent back-filling. T h e thickness and placing techniques for backfill concrete can have an effect upon the liner's subsequent ability to resist buckling. T h e concrete mix should be controlled to reduce the effects of heating the liner, which subsequently cools and pulls away from the concrete, and to limit shrinkage of the concrete. The introduction o f cold air or water to the liner also can cause the external gap to increase and, in conjunction with grouting or a buildup o f external water pressures, can result in buckling. After the liner has been in operation for a period of time, and depending on the deformation and subsequent recovery abilities of the rock surrounding the tunnel, a gap far in excess of that originally surrounding the liner may occur upon dewatering. This can be a critical time for buckling, especially if the operation of the facility has primed the surrounding ground with water at .pressures equal to the internal operating pressure. There are often circumstances in which it is not economically feasible to provide a plain steel liner sufficiently thick to resist the probable buckling

forces. Under such circumstances, the addition of external stiffeners over part or all of the length of the liner may be justified. However, any projections into the gap surrounding the liner should be detailed with caution, as they have a negative effect upon the subsequent placement of uniformly constituted concrete, and may require the excavation to be enlarged over and above that required to accommodate a slightly thicker, plain liner. The economics involved in deciding how best to resist buckling vary from project to project. There is no one rule that can be applied universally. Unless external pressure relief using drainage tunnels and drill holes is provided, it is customary to assume that the rock can be saturated up to the ground surface, as in Cases B and C above (Fig. 6). I f external drainage is used, it must be extensive in order to be effective, using measures described above, under "Length". The external pressures described in Case D may be significantly higher than those in Cases B or C. Such pressure may be induced by a nearby penstock, from pressures induced from seepage around the steel liner, or from an overlying reservoir. Evidence also indicates that highpressure seepage may occur along the penstock, induced from leakage around the end of the steel liner. This is particularly true for very high-head plants. Such leakage can bypass even the most effective external drainage system. This high-pressure water can build up a significant reservoir in the host rock and buckle the liner during a unwatering phase, unless unwatering is done very slowly, allowing drainage and equalization o f external pressures. To combat this potential problem, special treatment of the end section, the steel liner, may be necessary, as shown in Fig. 11. The combination of reinforcement, embedded collars and/or anchor rings, and high-pressure grouting is intended to limit access of high-pressure water behind the steel liner. For high-head schemes where such problems can occur, it may be more appropriate to design the steel liner for the full static head than to provide a complex drainage system. Some designers have used drainage pipes placed behind the steel liner, connected to the gap, to remove highpressure water on unwatering. This scheme should not be used; it is not effective in the long term, as the pipes can become clogged with precipitates.

Grouting
The following types of grouting are required over various portions of the tunnel: Contact Grouting--To fill large voids behind the steel and concrete liners that occur due to inadequate concreting,

168

TUNNELLING AND UNDERGROUNDSPACETECHNOLOGY

Volume 4, Number 2, 1989

STEEL SADDLE IF REINFORCEMENT CANNOT BE CARRIED THROUGH~ RADIAL AND LONGITUDINAL REIFORCEMENT~ \ \ \

L,~ ~ "

MULTI-LINE GROUT CURTAIN IANCNOR RINGS SEEPAGEBARRIERS) / STIFFENERS / ( SEEPAGEBARRIERS) ,,


,'.',' ".'" ".':...'.

/~( I
I / / [ /

." , ' . . . . = , :~.. .': =.'.. . . Z . ' ' "" ' ' - . . a . ~ =" L = . ~ .

I/i

__/[

L',,'.,'.'-."_e'

kx~STEEL LINER

cedent experience (Lysne 1971), rock traps may be sized to contain those rock pieces that can be transferred along the invert of the tunnel by the prevailing velocity and invert roughness. In general, it is good practice to concrete the tunnel invert for ease of inspection. The improved hydraulic efficiency of the tunnel allows easier transport of rock pieces and sand. Furthermore, much of the material derived from fully unlined tunnels comes from the invert, where erosion is concentrated.

Plugs
"P'p"'
,./' .~l,'.

j.',',

," " ? . .

:4

I
RING CONSOLIDATION GROUTING
=

I .....
l [

" , - ' . . . . . . :... : :., ...o...;.

..

I]
I

I I tI
t ]

GROUTINp RING CONSOLIDATION -

Figure I1. Seepage control at end of steel liner.


or due to air trapped during the concreting operations. Embedment Grouting--To seal the gap between the steel liner and concrete that forms due to concrete shrinkage, plastic set in the rock during loading/unloading, and the temperature differential between the liner and the mass rock. Consolidation Grouting--To consolidate blast-damaged or relaxed rock and to reduce leakage. Although some investigators have claimed improvement of the modulus of the rock as a result of consolidation grouting, this is a debatable point. Contact grouting, the first grouting process performed, is required only in the arch of the tunnel. In tunnels less than 6 m in diameter, a single line of grout spaced at 3-4 m normally is adequate. Stable mixes having a cement/water ratio of 1:1 (or thicker) by volume, with 0.5-1% bentonite, are appropriate. Modest pressures of 2-3 kg/cm 2 are adequate, although higher pressures are not detrimental. Embedment grouting of the gap between the steel and concrete liners normally is the second step. Grout mixes must be thinner, with about 2% bentonite to improve fluidity and penetration. Grouting should be done in rings (typically, six to eight holes per ring), moving upslope, with forward holes open to allow drainage. Pressures of up to 50% of the buckling pressure are appropriate to ensure against localized loading and buckling; however, the main requirement is to ensure gap-filling. If operation water is significantly colder than the rock, the liner can be cooled prior to grouting by spraying cold water on the liner to encourage the gap to form. The last step, consolidation grouting, is done by the ring method, moving upstream and grouting through the same holes as the embedment grouting. With the packer attached to the steel liner, the gap, as well as the fractured rock, is grouted a second time. It is advantageous to grout to the highest pressure possible without buckling the liner. Pressures of up to 75% of buckling pressure are advisable to prevent buckling due to localized overloading, because the grouting cannot be counted on to provide uniform, all-around pressure. Such grouting not only consolidates the rock, but also tends to induce a compressive stress in the concrete/steel liners. Such prestress, though desirable, cannot be counted upon to be maintained, because of rock creep. General criteria for consolidation and contact grouting of a tunnel lined only in concrete are the same as the criteria for a steel/concrete section.

Operational Aspects Debris Traps


Traps to collect soil and rock are necessary for unlined or shotcrete-lined tunnels, and in areas where debris c a n enter from the intake or surge shaft. The traps should be located upstream of the concrete or steel-lined portions. To reduce the first filling, it is advisable to pressure-wash the tunnel walls. Traps can be designed to remove most of the unsuspended soil load; however, the hydraulic aspects must be carefully studied. On the basis of pre-

Plugs must resist side shear and prevent excessive seepage around the plug. Differing design criteria are used by various individuals and agencies. Many plugs are sized simply to be about twice the tunnel diameter, independent of head and rock conditions Allowable shear stresses of 3-10 kg/cm~ are used, depending on the quality of the rock. Plugs should be designed such that their lengths satisfy criteria for acceptable shear strength and hydraulic gradients around the plug. These criteria are similar to those used for establishing the length of steel liner. However, because the openings where high-pressure water normally exits are smaller, hydraulic gradients may be greater. Special grouting and downstream pressure relief can be installed to control seepage and allow a reduction in length. Both contact and consolidation grouting are required. Contact grouting of short plugs can often be done by fanning holes from the rear end. However, for long plugs, embedded pipes with appropriate exit points for the grout, placed near the arch and carried to the downstream end, are effective for contact grouting. Consolidation grouting is often done by long inclined holes. However, long plugs may have a hollow downstream section, allowing access to a manway or hinged door. This arrangement permits ring grouting to be done from inside the hollow section, and improves the grouting seal. Such plugs may have to be reinforced along the hollow section. Plugs of this type can easily be extended or the grout curtain enhanced if excessive seepage occurs around the plug after watering of the tunnel. Large plugs require careful concrete control to prevent excessive heat of hydration and cracking. This procedure may involve cooling, especially if the schedule dictates rapid completion. Although a curing period of 28 days is advisable prior to grouting, it c a n be shortened if necessary. Rings of drainage holes drilled into the rock at the downstream end of the plug are advisable to ensure stability of the tunnel.

Volume 4, Number 2, 1989

TUNNELLINGAND UNDERGROUNDSPACETECHNOLOGY 169

In the absence o f general design criteria for the sizing o f plugs, the values listed in Table 7 are suggested.

Watering/Unwatering
T h e initial filling o f the pressure tunnel should be carefully controlled to limit cracking o f the concrete liner. Such cracking is intensified d u e to the difference in pressure between the g r o u n d w a t e r and tunnel water. Normally the tunnel has been o p e n for several years d u e to construction, and drainage o f the rock has occurred. Filling the tunnel slowly allows pressure equalization to occur, and thereby limits d e f o r m a t i o n o f the rock and liners. As r e c o m m e n d e d by Deere (1983), at least two to three weeks generally should be allowed for this process; however, this time period may be excessive for relatively low-head plants. A filling rate o f 5 to 10 m/h normally is adequate. I f possible, monitoring water levels in the surge shaft with a closed u p s t r e a m gate will allow a good estimate o f leakage. Such leakage usually is high d u r i n g filling, as the g r o u n d w a t e r table is re-established. I f leakage accelerates at a specific internal pressure, hydraulic jacking or piping may be occurring. Piezometers should be placed in areas where significant construction drainage is expected, a n d should be read as filling takes place. Such piezometers are best installed in exploratory holes used for pre-design investigations, so that the history o f g r o u n d water changes is known.

In addition to g r o u n d w a t e r monitoring, m a p p i n g o f existing springs, seepages and landslides should be done before watering. Changes in these conditions should be noted by a regular inspection p r o g r a m . A survey of cracks in the concrete should be done prior to first filling. These cracks should be m a p p e d and, if unwatering is planned, spray-painted so that extensions or new cracks can be identified upon unwatering. Unwatering o f pressure tunnels should also be done carefully at a rate between 2 to 10 m/hr, utilizing slower rates for high-head plants. G r o u n d water changes should be noted as the unwatering takes place. Especially for very high-head projects, significant reservoirs can be built up in the rock mass. I f drawdown is too rapid, high pressures will be exerted on the tunnel linings. D e p e n d i n g on the strength and buckling resistance o f the various elements, local failures can be initiated. A detailed inspection o f the tunnel should be d o n e immediately after unwatering is complete. Records o f inflow, local failures o f the rock or lining, cracking o r other distress should be recorded. []

References
Amstutz, E. 1970. Buckling of pressure shaft and tunnel linings. Water Power November 1970. Bieniawski, Z. T. 1974. Geomechanics classification of rock masses and its application in tunnelling. In Proc. Third lnt. Congr. o~ Rock Mechanics, Vol. 11A. Denver': International Society of Rock Mechanics. Brekke, T. L. and Ripley, B. D. 1987. Design guidelines for pressure tunnels and shafts. EPRI Document RE- 1745-17. Broch, E. 1984. Unlined high pressure tunnels in areas of complex topography. Water Pwr Dam Construction 36 (11), 21-23. Deere, D. U. 1983. Unique geotechnical problems at some hydroelectric projects. In Seventh Pan Am Proc., Vancouver, pp. 865-88. Canadian Geotechnical Society. Deere, D. U., Peck, R. B., Parker, H. W., Monsees, J. F. and Schmidt, B. 1900. Design of tunnel support systems. Highway Research Record 339. Hendron, A..]., Fernandez, G. and Lenzini, P. A. 1987. Design of pressure tunnels. In Symposium on the Art and Science of Geotechnical Engineering, Vol. 1. Englewood Cliffs, N.J.: Prentice-Hall. LamC G. 1852. Lemons sur la th~orie math~matique de l'Elasticit& Vol. I. Paris: Bachelier. Lysne, D. K. 1971. Sand transport and sand traps in hydro power tunnels. In Proc. Pumped Storage Development and its Environmental Effects, University of Wisconsin, September, 1971. Milwaukee, WI: American Water Resources Association. Merritt, A. H. 1972. Geological prediction for underground excavation. In Proc. First North American Rapid Excavation and Tunneling Conf., pp. 115-132. New York: A1ME. Sawatsky, L. Personal communication, October 1986. Singh, B., Nayak, G. C. and Kumar, R. 1988. Design recommendations for plain concrete lining in power tunnels. In Proc. Int. Symp. on Tunnelling for Water Resources and Power Projects, Vol. 1. New Delhi: Central Board of Irrigation and Power. Solvik, O. 1984. Unlined tunnel hydraulics. Hard Rock Engineering. Oslo: FHS. Terzaghi, K. Rock Defects and Loads in Tunnel Supports in Rock Tunnelling with Steel Supports (R. V. Proctor and T. White, eds) Youngstown, OH: Commercial Shearing and Stamping Co. Walia, M. 1985. On empirical classification systems for support pressures estimation for underground structures. Paper presented at Tunnelling Association of Canada Conference, August 1985. Wickham, G. E., Tiedemann, H. R. and Skinner, E. H. 1972. Support determination based on geological predictions. In Proc. First North American Rapid Excavation and Tunneling Conf., Vol. 1, pp. 43-64. New York: AIME.

Acknowledgements
T h e writer thanks Dr A. Merritt, Dr L. Jory, Mr I. Pinkerton and Mr C. H. M a a r t m a n for their review o f this paper, and their very constructive comments.

Table 7. Acceptable hydraulic gradients for plug length.

General rock type


Massive, hard, widely jointed Hard to moderately h a r d , moderately jointed Moderate to weak, moderately j o i n t e d W e a k closely jointed or sheared V e r y weak, possibly erodible

Maximum shear stress (kg/cm 2)


5 3 2 1 0.5

Maximum hydraulic gradient


15 - 2 0 10 - 14 7 - 9 5 - 6 3 - 4

Notes: Erodible features downstream of plug to be treated locally.

170

TUNNELLING AND UNDERGROUND SPACE TECHNOLOGY

Volume 4, N u m b e r 2, 1989

Anda mungkin juga menyukai